Publication 3: Are mouse models suitable to study human ageing?

The translatability of mouse ageing to human ageing was investigated using urinary proteome analysis. The urinary proteome in 89 wild-type (C57/BL6) mice aged between 8-96 weeks was analysed using CE-MS. Using age as a continuous variable, a correlation analysis was performed with age and 163 significantly age-correlated peptides in mice were identified (Table S5.1). The peptides predominantly included several collagen uromodulin fragments and other mouse-specific peptides. An orthology analysis was conducted comparing the 163 age-correlated peptides identified in mice with the 116 age-correlated peptides identified in

6 Summary of the results 79 the 1227 healthy subjects used from the previous human ageing study. 54 unique peptides including collagen alpha-1(I) chain, collagen alpha-1(III) chain, collagen alpha-2(I) chain and uromodulin (Table S5.2) in mice showed homology to 40 unique sequences in humans. Out of the homologous sequences, a 100% homology was detected with 13 human peptides. These peptides included predominantly collagen alpha-1(I) chain (n=11) (Figure 5.1).

Figure 5.1: Comparison of orthology in collagen alpha-1(I) chain in mouse and human

To further investigate whether the mouse urinary peptides were representative of human ageing we developed multidimensional models based on the ortholog peptides and scored mice and human age using these models based on the hypothesis that correct age classification by these ortholog peptides in both mice and humans validates the translatability of the mouse peptides. Thus, an ageing classifier called ACM54 (ageing classifier in mouse 54) was developed and validated using the ortholog age-correlated peptides (N= 54) in mice using a training cohort of wild-type mice (N= 39) and a test cohort of wild-type including young (12 weeks; N= 15), mature (48 weeks; N= 15) and old (96 weeks; N= 15) in a support vector machine (SVM)-based modelling. In this independent validation the ACM54 classifier was successfully able to discriminate between the different age groups (Figure 5.2).

6.4 References

(1) Levey AS, Coresh J. Chronic kidney disease. Lancet. 2012;379(9811):165-180.

7 Discussion 83

7 Discussion

Molecular mechanisms leading to ageing are still under investigation. The aim of this thesis was to identify age-related biomarkers in the hope to unravel additional molecular events associated with ageing.

In a preliminary study, different “omics” approaches were investigated in chronic kidney disease (CKD) patients in order to select the appropriate method to study ageing. CKD is a leading public health problem and it generally comprises a heterogeneous group of diseases affecting kidney structure and function (1). Globally, the prevalence of CKD is over 10%

widely affecting over 70 years old individuals (2). There are five different stages of CKD starting with a normal stage characterised by a normal renal function and a final stage defined as end-stage renal disease (ESRD) characterised by an irreversible loss of renal function (1).

Finally, ESRD is usually followed by death except in the case of renal replacement therapies like transplantation or dialysis.

The performance of proteomic and metabolomic techniques was compared in urine and plasma in the management of CKD to evaluate which of the approaches can be used in further ageing studies. In comparison to other “Omics” technologies, metabolomics is the study of metabolites. Metabolites are small molecules generated in metabolic reactions by the enzymes of the cell (3). As the metabolite patterns of a cell or an organism reflect gene expression, they can also be considered to closely reflect cellular functions in comparison to genes and proteins. However, due to chemical complexity, high variability of metabolites and lack of standardised protocols for metabolomic analyses, metabolomic patterns are less explored.

Findings described herein demonstrated that urinary proteomics, urinary metabolomics and plasma metabolomics were both efficient technologies in the diagnosis and prediction of renal damage even though urinary proteomics showed a slightly better performance in the prediction analysis. Furthermore, there was no added value in the diagnosis and prognosis of CKD in combining urinary proteomics and plasma metabolomics. In conclusion, not only urine contains necessary information to investigate CKD but also urinary proteomics is as powerful as urinary and plasma metabolomics in the management of CKD. Urinary proteomics was therefore established as a powerful tool to investigate an age-related disease and by extrapolation also ageing.

Since urinary proteomics was established as an appropriate technique, I then used it for further ageing studies. In a unique cohort of 11560 individuals, the proteome was investigated.

Through this unique cohort, the comparison between molecular events occurring during

normal ageing and pathological ageing was for the first time investigated using urinary proteome analysis. Normal ageing would be defined here as the normal process of ageing where as pathological ageing would be ageing observed as a consequence of disease.

Results depicted perturbations in collagen homeostasis and trafficking of toll-like receptors and endosomal pathways associated with both normal and pathological ageing. In addition, perturbation in the insulin-like growth (IGF) factor pathway was only observed in pathological ageing.

Though alteration in collagen homeostasis was not mentioned as an important molecular event occurring during ageing the well-known review summarising the hallmarks of ageing (4), in the urine, it appears to be of value. Collagen homeostasis is crucial during the development and perturbations or alterations in collage homeostasis lead to several conditions including fibrosis, cancers and cardiovascular conditions (5). Moreover, identification of trafficking of toll-like receptors (TLRs) and endosomal pathways associated with ageing depict a perturbation in the immune system caused by ageing and specifically the concept of

“inflamm-ageing”. It has been speculated that perturbations in TLRs pathway can cause imbalance in inflammation (6). Inflamm-ageing was recently proposed as a theory leading to ageing. It is believed to be the result of the accumulation of antigenic exposure throughout years causing inflammatory responses and eventually leading to tissue damage (7).

Surprisingly, the IGF pathway was shown to be affected during ageing via urinary proteome analysis. The IGF pathway has been extensively studied in animal models including Caenorhabditis elegans and mammals (8) and to a less extent in humans. Being able to detect perturbations in the IGF pathway in urine highlight urine as a relevant biological fluid in the study of ageing when used in combination with appropriate tools.

Finally, to investigate the translatability of mouse ageing to human ageing and achieve a wholesome picture on ageing, ageing biomarkers were also identified in wild-type mice and compared with ageing biomarkers in healthy humans. Ortholog peptides common in both mice and humans, depicted perturbations in collagen homeostasis as a key molecular change observed during ageing. Furthermore, ageing classifiers established on these ortholog peptides were able to discriminate the age in both wild-type mice and healthy subjects. Thus, suggesting that focussing on urinary peptides, mouse ageing can be translated to human ageing. Therefore, research can utilise mouse models for the evaluation of intervention strategies for the management of age-related complications in humans.

7 Discussion 85

7.1 References

(1) Levey AS, Coresh J. Chronic kidney disease. Lancet. 2012;379(9811):165-180.

(2) Levey AS, Inker LA, Coresh J. Chronic Kidney Disease in Older People. JAMA.

2015;314(6):557-558.

(3) Goodacre R, Broadhurst D, Smilde AK et al. Proposed minimum reporting standards for data analysis in metabolomics. Metabolomics. 2007;3(3):231-241.

(4) Lopez-Otin C, Blasco MA, Partridge L, Serrano M, Kroemer G. The hallmarks of aging.

Cell. 2013;153(6):1194-1217.

(5) Bonnans C, Chou J, Werb Z. Remodelling the extracellular matrix in development and disease. Nat Rev Mol Cell Biol. 2014;15(12):786-801.

(6) Olivieri F, Rippo MR, Prattichizzo F et al. Toll like receptor signaling in

"inflammaging": microRNA as new players. Immun Ageing. 2013;10(1):11.

(7) Baylis D, Bartlett DB, Patel HP, Roberts HC. Understanding how we age: insights into inflammaging. Longev Healthspan. 2013;2(1):8.

(8) van HD. Insulin, IGF-1 and longevity. Aging Dis. 2010;1(2):147-157.